Auxiliary Armed Forces and Innovations in Security Governance in Mozambique’s Civil War

Abstract

Who rules during the civil war? This article argues that the concept of armed group governance must be expanded to include auxiliary armed forces linked to rebels or the government. Comparing the organization of rebel and government auxiliaries, the article demonstrates that security governance during war is never static, but evolves over time. Evidence from the civil war in Mozambique (1976–1992) shows that the auxiliary’s origin shapes its initial level of autonomy. Second, auxiliary contribution to battlefield success of one side may induce innovations adopted by auxiliaries on the other. Both have distinct consequences for the nature of governance.

Full text

Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=fciv20
Civil Wars
ISSN: 1369-8249 (Print) 1743-968X (Online) Journal homepage: http://www.tandfonline.com/loi/fciv20
Auxiliary Armed Forces and Innovations in Security
Governance in Mozambique’s Civil War
Corinna Jentzsch
To cite this article: Corinna Jentzsch (2017) Auxiliary Armed Forces and Innovations
in Security Governance in Mozambique’s Civil War, Civil Wars, 19:3, 325-347, DOI:
10.1080/13698249.2017.1412752
To link to this article: https://doi.org/10.1080/13698249.2017.1412752
© 2017 The Author(s). Published by Informa
UK Limited, trading as Taylor & Francis
Group
Published online: 19 Jan 2018.
Submit your article to this journal
View related articles
View Crossmark data

CIVIL WARS, 2017
VOL. 19, NO. 3, 325347
https://doi.org/10.1080/13698249.2017.1412752
Auxiliary Armed Forces and Innovations in Security
Governance in Mozambique’s Civil War
CorinnaJentzsch
Institute of Political Science, Leiden University, Leiden, The Netherlands
ABSTRACT
Who rules during the civil war? This article argues that the concept of armed group
governance must be expanded to include auxiliary armed forces linked to rebels or
the government. Comparing the organization of rebel and government auxiliaries,
the article demonstrates that security governance during war is never static, but
evolves over time. Evidence from the civil war in Mozambique (1976–1992) shows
that the auxiliary’s origin shapes its initial level of autonomy. Second, auxiliary
contribution to battleeld success of one side may induce innovations adopted
by auxiliaries on the other. Both have distinct consequences for the nature of
governance.
Introduction
Who rules during civil war? Given the limits of many states to address security
threats and other governance challenges under wartime conditions, alternative,
localised arrangements often emerge through which those in power coordinate
the implementation of policies with other actors, delegate tasks to them or
merely tolerate their involvement in service provision and conict resolution
(Arjona 2014, Staniland 2012, 2017). This multi-actor, multi-layered nature of
wartime governance1 empowers non-state actors and nurtures opportunistic
political and economic interests, which may pose signicant challenges to state-
and peacebuilding (Seymour 2014).
A good illustration of this dynamic is the case of Somalia. In response to the
absence of a central government, ‘informal systems of adaptation, security, and
governance’ emerged in local communities that tried to ‘adapt in a variety of
ways to minimise risk and increase predictability in their dangerous environ-
ments’ (Menkhaus 2007, pp. 74, 75). Business leaders, civil society organisations,
© 2017 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-
NoDerivatives License (http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial re-use,
distribution, and reproduction in any medium, provided the original work is properly cited, and is not altered, transformed,
or built upon in any way.
CONTACT Corinna Jentzsch c.jentzsch@fsw.leidenuniv.nl
OPEN ACCESS

326 C. JENTZSCH
and other (armed) actors develop an interest in prolonged warfare, lawlessness,
or state failure and, as a consequence, act as potential ‘spoilers’ in the peace
process. The way in which those in power nurture, respond to, coordinate or
undermine such alternative governance systems is crucial to understanding
successes and failures in state- and peacebuilding.
Struggle in the domain of security governance is the focus of this article. By
security governance, I mean arrangements between and amongst state and non-
state (armed) groups to maintain territorial control. Such arrangements entail a
‘fragmentation of power and authority’ in the security realm (Krahmann 2003, p.
20) and devolution of responsibilities within a network of actors (Abrahamsen
and Williams 2008, p. 545).
Research on fragile statehood and civil war has explored the fragmentation
of armed groups (Sinno 2008, Bakke et al. 2012), the emergence of alliances
between armed actors to inuence war outcomes (Christia 2012, Seymour 2014),
‘wartime institutions’ that govern civilians (Arjona 2014) or ‘armed politics’ that
shape how states and armed groups interact (Staniland 2017). I build on this
research to analyse change in alliances and their implication for security gov-
ernance. The fragmentation of authority in the security realm is often shaped by
what state and non-state actors learn from their opponents on how to secure
territorial control. However, empowering armed actors to assist with security
governance may intensify the conict by inciting more violence (Clayton and
Thomson 2014), which in turn may militarise security governance further. This
dynamic is the focus of this article. I show that even though initial conditions
inuence how security governance is organised, they evolve through processes
of experimentation, innovation, and learning that largely respond to dynamics
on the battleeld; hence the notion that governance during civil war is domi-
nated by a military logic (Wood 2008).
Auxiliary armed forces play a crucial part in that innovation process. These
groups consist of civilians with little to no military training and limited access to
weapons. They are frequently referred to as self-defense forces, vigilantes and
militias that are either aligned with the government or the rebels, or attempt to
remain independent. I use the term ‘auxiliary armed forces’ for two reasons. First,
it is neutral enough to include auxiliaries both on the government and rebel
side. In contrast, previous work denes ‘militias’ as counter-insurgent armed
groups who emerge to protect the local population from rebels (Jentzsch 2014).
Second, the term ‘auxiliary’ emphasises that such forces assist the rebels or the
government in their eorts to control the local population. The kind of relation-
ship and division of labour that emerges between the principal group and the
auxiliary inuences wartime governance.
In this article, I compare the organisation and evolution of auxiliary armed
forces on the rebel side with those on the government side in order to demon-
strate that security governance during war is never static, but evolves over time.

CIVIL WARS 327
Using evidence from the civil war in Mozambique (1976–92), I contrast the align-
ment between state armed forces and auxiliaries with that between rebels and
auxiliary armed forces. Both the Renamo (Resistência Nacional de Moçambique)
rebels and the Frelimo (Frente de Libertação de Moçambique) government aligned
with auxiliary armed forces to function as a liaison to the civilian population and
support the ght against the other. Renamo created the mujeeba in the central
and northern regions as a separate force from regular combatants. The mujeeba
lived amongst the population and were responsible for vigilance, collecting food
from the population for the base and enforcing rebel rule. In 1988–89, Frelimo
aligned with a grassroots movement in the Centre and North, the Naparama,
which had emerged independently amongst the displaced. Naparama took
over policing and military tasks. The Naparama posed such a severe threat to
Renamo that Renamo elites decided to form a stronger auxiliary armed force (an
‘enhanced mujeeba’ force) in 1990–91. Its role was more oensive than that of
the prior mujeeba and focused on defeating the Naparama. With this renewed
eort, Renamo managed to gain the upper hand again and made signicant mil-
itary advances in central and northern Mozambique, shortly before the signing
of the peace agreement in October 1992 that ended the war. This article traces
this learning process and points to critical junctures that inuenced innovations
in these multi-actor, multi-layered governance systems and their consequences
for governance outcomes.
The article proceeds as follows. I rst provide a short introduction to political
order in civil wars, rebel governance and militias. I then analyse Renamo’s rela-
tionship with civilians and the role of auxiliary armed forces in security govern-
ance in rebel-held zones. I discuss how civilians in the government-held zones
responded to Renamo’s threat, and in turn how Renamo reacted to increased
pressure from civilian-based militias. The evidence for this paper comes
from interview and archival data, collected during 14months of eldwork in
Mozambique between 2010 and 2016, and secondary sources.2
Security Governance and Auxiliary Armed Forces in Civil War
Auxiliary armed forces are crucial for both the state and armed groups to create
links between them and the local population they control. First, civil war does
not necessarily create disorder, and violence is not the only way rebels interact
with civilians (Weinstein 2007, Mampilly 2011, Arjona et al. 2015, Arjona 2016,
Huang 2016, Terpstra and Frerks in this issue). Rebel groups (re-)create norms
and institutions to optimise their control over the local population and occupied
territory, gain access to resources and in most cases, especially in secessionist
wars, govern (Reno 1998, Jackson 2003, Mampilly 2011, Kasr 2015, Malejacq
2016). Rebels’ relations with civilians vary with respect to the degree to which
civilians are involved in the administrative and decision-making processes, how

328 C. JENTZSCH
responsive armed groups are to civilian preferences, and to what degree armed
groups regulate civilian life (Kasr 2015, Weinstein 2007, Arjona 2016, Terpstra
and Frerks in this issue). While some armed groups take people’s needs into
account and provide extensive public services to create their own quasi-state,
such as the LTTE in Sri Lanka (Mampilly 2011, Terpstra and Frerks in this issue),
others limit their interactions with civilians and focus on resource extraction,
such as Renamo in Mozambique (Vines 1991, Weinstein 2007).
In both cases, once rebels occupy territory, they need to engage with the
local population in order to solidify territorial control (Weinstein 2007, p. 164,
Kalyvas 2006, p. 107). In Sri Lanka, for example, the LTTE was supported by
three civilian auxiliary units of several thousand men that were recruited from
the local peasant population (Richards 2014, p. 29). In South Sudan, the White
Army, a loose network of self-defense units composed of civilian youths, has
in recent years provided a liaison for the SPLA-in-Opposition, led at the time
by Riek Machar (Breidlid and Arensen 2014). Auxiliaries of the FARC and ELN in
Colombia organised the population in urban areas, with a special focus on the
political education of the masses (Dieterich 2016, Velasquez 1995). These auxilia-
ries shape rebel interactions with civilians, as they help to solve the ‘identication
problem’ (Kalyvas 2006, p. 107). In irregular war, rebels (and their supporters)
tend to hide amongst the population. Local intelligence is necessary to distin-
guish between supporters and defectors and can be provided by auxiliaries
with close contacts with civilians. Despite their crucial function in the internal
organisation of rebel groups and systems of rebel governance, rebels’ auxiliary
armed forces have so far received little attention in scholarly research. Most
recent work on armed group fragmentation focuses on competition between
dierent armed groups or units on the rebel side (Bakke et al. 2012, Christia
2012, Sinno 2008). In contrast, this article analyses how dierent units within
the same organisation complement each other.
Research on auxiliaries on the government side has increased over the last
few years, with signicant advances in identifying the causes and consequences
of their formation and collaboration with state security forces (see e.g., Carey
et al. 2013, Schubiger 2013, Mitchell et al. 2014, Clayton and Thomson 2014, Peic
2014, Jentzsch 2014, Jentzsch et al. 2015, Biberman 2016). Within the broader
framework of ‘armed politics’ (Staniland 2017), state auxiliaries are those armed
groups with which states align themselves. States can either mobilise such forces
themselves (top-down mobilisation) or co-opt existing forces (bottom-up mobi-
lisation) (Jentzsch et al. 2015). In Sierra Leone, for example, the Kamajor formed
as local defense forces, which were later co-opted by political elites and became
a professional armed force that even substituted for the state army (Homan
2011). Governments frequently delegate tasks to such forces to increase access
to intelligence, multiply manpower, and/or avoid accountability (Carey et al.
2013, Mitchell et al. 2014). In Sudan’s Darfur region, the government outsourced

CIVIL WARS 329
violence against the rebellion to local Janjaweed militias to avoid being held
accountable (De Waal 2004).
Auxiliary armed forces can have defensive, oensive or purely logistic pur-
poses. Their primary purpose is to provide assistance in population control, infor-
mation gathering, resource extraction and logistics to their respective ‘senior’
partner. I argue that two factors – a structural and a dynamic one – inuence the
division of labour between auxiliary and senior partner in particular, and secu-
rity governance more broadly. First is the origin of the auxiliary. If the auxiliary
emerges independently of the senior partner, and cooperation between the
two is ad hoc, auxiliaries tend to have more autonomy and take on a variety of
tasks. If the senior partner is involved in the formation of the auxiliary, respon-
sibilities tend to be limited and clearly dened, as they full concrete tasks in a
well-organised system of governance.
Second, given the military context in which governance emerges, successes
and failures on the battleeld may lead to ‘tactical innovation’ and ‘tactical adap-
tation’, which may initiate auxiliary armed group formation and co-optation in
the rst place, or increase the auxiliaries’ autonomy, and change and expand
their activities as a consequence.3 This process is similar to what scholars in
policy research have labelled ‘learning’, the updating of previously held beliefs
to make policies more eective (Dobbin et al. 2007). Armed groups can learn
from a rival’s tactics and adopt them to improve their own system of security
governance. This dynamic has implications for governance outcomes for non-
combatants, as protection and the enforcement of rules may become more
militarised and violent if armed groups are more involved in the governance
of civilians (Wood 2008).
I develop and explore these arguments with evidence from the civil war in
Mozambique (1976–92). The advantage of this case is that the conict had two
main parties – the incumbent versus the insurgent. Over the course of the war,
both sides developed similar tactics and came to resemble each other in peo-
ple’s post-war narratives (Bertelsen 2016, pp. 73, 74, 83). Crucially, both on the
rebel and the government side, auxiliary armed forces supported governance
eorts. I focus on the war in the central and northern provinces of Nampula
and Zambézia where Renamo was more involved in governance and its vio-
lence tended to be less atrocious than in the South, where Frelimo aligned with
grassroots auxiliaries, and where the interaction between the auxiliaries of the
two sides can be well observed.4 I specically focus on the eect of auxiliaries
on governance outcomes; violence perpetrated by Renamo and Frelimo com-
batants lie outside the scope of this article.5

330 C. JENTZSCH
Security Governance in Mozambique’s Civil War
Auxiliary Armed Forces in Rebel-held Areas
Renamo was formed with the help of the Rhodesian intelligence service in the
late 1970s (Vines 1991). Armed activity began in the centre of Mozambique, but
then extended, over the course of the 1980s and with the support of Apartheid
South Africa, to the south and north of the country. The overall goal of Renamo
attacks was to reverse Frelimo’s socialist policies after independence by destroy-
ing state infrastructure and communal villages – a prime symbol of Frelimo’s
socialist vision of society, economy and government (Vines 1991, Coelho 1998).
When the rebels began to occupy rural areas, they reversed Frelimo’s secularist
policies, as this respondent outlined: ‘They told us: if there’re Christians [among
you], then worship, if there are chiefs [among you], then pray to the spirits […]
and conduct [your] ceremonies’.6
Despite these political goals, Renamo made limited attempts to mobilise
the people and recruit volunteers, though there was some variation in political
indoctrination across provinces (Schafer 2007, p. 66). Discontent with Frelimo
by marginalised traditional leaders and the rural peasantry and local conicts
increased support for Renamo (Geray 1990, Lubkemann 2005, Roesch 1989).7
However, Renamo’s main strategy of recruitment was the abduction of young
men, including many children (Hanlon 1984, p. 229, Schafer 2007, pp. 58, 68,
Bertelsen 2016, p. 30).8 Promises of economic and political benets, threats
of punishment and a process of ‘resocialization’ convinced the abducted to
stay with the rebels (Vines 1991, p. 95, Schafer 2007). Traditional leaders often
decided to side with Renamo for the same reasons (Roesch 1992, Geray 1990,
Pereira 1999b).
The rebels’ limited political structure and eorts of mobilisation led ana-
lysts to conclude that ‘Renamo was rst and foremost a military organization’
(Finnegan 1992, p. 74). The nature of the group’s organisation supported this
claim. While Renamo’s political organisation was relatively weak, its military
organisation was strong. The organisation had a centralised military hierarchy,
which was supported by South Africa’s supply of a sophisticated radio network
(Vines 1991, p. 82).
9
Afonso Dhlakama was the commander-in-chief, assisted by
a 15-member military council composed of three chiefs-of-sta for the northern,
centre and southern zones, 10 provincial commanders, and Dhlakama’s personal
sta. Provinces were subdivided into two to three regional commands. One
regional command consisted of a brigade, which consisted of several battal-
ions (each about 250 men), companies (100–150 men), platoons (30 men) and
sections (10 men).10
The construction of Renamo bases reected the group’s centralised military
hierarchy. In the central and northern regions, where discontent with Frelimo
was higher, Renamo was able to establish more permanent bases. Renamo’s
bases in the South, where support for Frelimo was strong, were more mobile

CIVIL WARS 331
(Roesch 1992). Amongst these bases were ‘regional’ or ‘provincial’ bases, in which
the commander of that geographical unit resided. Casa Banana in Gorongosa
district in Sofala province was Renamo’s headquarters until the Zimbabwean
re-capture of the base in 1986.
The areas in which Renamo established military bases were part of the
rebels’ ‘control zones’, in contrast to ‘tax zones’ and ‘destruction zones’ (Gersony
1988). In Kalyvas (2006) framework, these zones correspond to areas under
full, partial or contested territorial control. Control zones were areas in which
the population was involved in food production for Renamo and assisted in
the transport of supplies to the base. ‘Tax zones’ were areas in which Renamo
combatants collected food contributions from the population and abducted
people. ‘Destruction zones’ experienced frequent Renamo attacks until they
were completely destroyed and their residents had ed.
Governance was most common in control zones, though Renamo’s provision
of services in exchange for supplies remained limited (Gersony 1988). Some
of the former Renamo combatants I spoke with had been trained as nurses,
but they appear to have treated combatants only.11 Some refer to schools that
existed within areas under Renamo control,12 but as one respondent said, ‘they
were constantly burned down’, presumably by Frelimo forces.13 As Vines (1991,
p. 93) argues, the only reciprocity that Renamo oered was ‘religious tolerance,
access to ICRC emergency food aid and the opportunity to remain alive’. Military
priorities would always override concessions to civilians. This is why Weinstein
(2007, p. 186) states that Renamo governance was ‘unilateral’ and relied on ‘lim-
ited participation’. He found little evidence that noncombatants were involved in
political decision-making and demonstrates that governance benetted mostly
the armed organisation, and not the civilians living under its control.
The evidence I collected leads me to a similar conclusion, despite the exist-
ence of a system of indirect rule through civilian administrators and traditional
authorities (Geray 1990, Pereira 1999b). Below Renamo’s military hierarchy
existed an administrative hierarchy that included traditional authorities, mam-
bos and their assistants, mujeeba, (also called mujuba, majiba, madjuba or
madjuhba).14 The empowerment of traditional authorities implied a signicant
recognition of their power to settle community conicts and regulate daily
life (Geray 1990, p. 118). In the words of a mambo in Nampula, they were in
charge of ‘the health of the family and the community’.15 The mambos’ tasks
included the redistribution of the population in the area held by Renamo, the
establishment of advance posts, the maintenance of surveillance and intelli-
gence networks, and the recruitment of local auxiliaries, the mujeeba, and, in
the North, a local police force, which Geray refers to as capeceiros (Geray
1990, p. 119).16 My respondents spoke of civilian administrators (‘delegados’)
in Renamo strongholds such as Namarrói in Zambézia who were tasked with
counting and controlling the population, while mambos focused on dispute
resolution and the collection of food for the base.17

332 C. JENTZSCH
The involvement of civilians and empowerment of traditional authorities
represented an attempt to legitimise Renamo rule. At the same time, however, it
implied a strict separation between civilians and combatants, which was ensured
through the geographical location of military bases away from the population
and the rigid regulation of combatants’ interactions with civilians. Military bases
were situated in deep forests and close to rivers for water supply.
18
Control posts
limited access to the centre of the base where the main commander resided.19
The civilian population did not have access to the base, but lived in concentric
circles around it, thus serving as a disguise for the base, a ‘human shield’, and
informers (Geray 1990, Vines 1991, p. 91).20 When civilians sought to contact
combatants, they went to the nearest control post (Geray 1990, p. 118). Strict
rules regulated combatants’ interactions with civilians (Schafer 2007, p. 70)
and combatants’ movements were closely controlled.21
In order to keep close control over the population without being engaged in
policing themselves, Renamo formed a separate force for that task, the mujeeba
(‘informers’). Mujeebas were responsible for the collection of information, food
and taxes from the population.22 Mujeebas were clearly distinguished from reg-
ular combatants. They had limited access to weapons and training and were
only allowed to carry cutting instruments (in contrast to combatants’ access
to AK-47s) (Vines 1991, p. 92). Renamo ocials did not consider them part of
the group’s ghting forces and did not completely trust them. While regular
combatants often fought in areas far away from their homes (Wiegink 2013, p.
115), mujeebas lived amongst the population in their area of origin. They were
not allowed to come close to the base as, in case they ed to government-held
areas, they might take Renamo ‘secrets’ with them (Geray 1990, pp. 117, 118).23
Mambos recruited mujeebas, voluntarily or by coercion, from local youth or from
amongst those captured during raids (Vines 1991, p. 92).24
The mujeeba’s role as Renamo’s local intelligence and enforcement agents
created an oppressive form of governance. In the rst public account of the
mujeeba in 1985, Renamo Secretary General Fernandes considered their access
to information and local knowledge as the most crucial characteristics: ‘The
mujeeba is our representative at the village level. He knows everybody in his
village. Nobody can come without being known. Then nobody also can betray
us because he surveys the area’.25 A former Renamo combatant in Nampula
showed that the mujeebas’ role was that of a vicious secret police and strict
enforcer of Renamo rule:
The work of the mujeeba among the people was mainly the work of secret police
agents. Where people would go, to church, [the mujeeba] would be there. In order
to control the population and those who might be against Renamo. Those [who
were against Renamo] they arrested and took to the base. And there were also peo-
ple in the neighborhoods that seemed [mentally ill] and spoke badly of Renamo;
the base sent mujeeba there as well. To arrest them. They were all arrested. And
when the combatants went to the bush to stage attacks, [the mujeeba] carried
material and assisted the combatants. They were also sent to collect food from

CIVIL WARS 333
the population to bring to the base. They went to search for chickens, [and often
took them] by force. And pigs. That’s how it was. And women, and everything.
To take to the base. So these were assistants to the base. They were the most
dangerous people.26
The excerpt shows that the primary task of mujeebas was to ensure discipline,
arrest any inltrators and prevent people from eeing.27 In cases in which people
attempted to escape, mujeebas were tasked to search and kill the defectors
(Hall 1990, p. 57, Vines 1991, p. 93).28 As the excerpt shows, mujeebas also col-
lected food from peasants to bring to the base (Gersony 1988, p. 24, Vines 1991,
pp. 92, 93, Pereira 1999a, p. 45). Contrary to what Secretary General Fernandes
claimed in 1985 in the same interview quoted above, food collection was largely
a coercive endeavour; some respondents report that those who refused to pro-
vide food to the mujeebas would be killed.29 Mujeebas frequently abused their
powers and took not only food, but also women by force.30 Respondents who
lived in government-held areas during the war claimed that it was the mujeebas
who did the killing, pillaging and raping in the villages, not the regular Renamo
combatants.31
Mujeebas also went on missions to inltrate government-held areas to collect
information about troop movements and planned operations.32 As a conse-
quence, Frelimo used the term ‘mujeeba’ for any displaced people arriving in
government-held areas, labelling them as potential Renamo collaborators.33
Accusing the displaced of spying gave Frelimo troops a reason to punish them,
or at least prevent them from entering Frelimo-held areas to seek refuge (Lemia
2001, pp. 47, 48). In a war in which it was dicult to identify who belonged to the
other side, peasants in contested areas were considered potential collaborators,
‘mujeebas’ (Vines 1991, p. 99).
Overall, Renamo’s delegation of tasks to auxiliary forces represented its strict
separation of military from social and political aairs, and served its major stra-
tegic and military interest to enforce control over the population and occupied
territory. Renamo did attempt to legitimate its war in the centre and north by
appealing to traditional sources of power. Re-instating traditional authorities
and mobilising traditional spirit mediums to support Renamo combatants were
two strategies in this regard (Wilson 1992). I agree with Weinstein (2007, p. 182),
though, when he concludes,
Although the reappointment of régulos [traditional authorities] was an eective
political appeal, it was also a strategy aimed at centralizing political control in a
hierarchical structure and limiting both the participation of civilians and their
capacity to shape the trajectory of the organization.
Furthermore, Renamo appealed to traditional religion in order to develop
a ‘cult of military prowess’ to further military interests rather than forge a link
between combatants and the local population (Wilson 1992). Renamo did
not have a monopoly on ritual sources of power and was soon challenged by
Frelimo’s auxiliary armed forces that also appealed to traditional religion.

334 C. JENTZSCH
Governance in rebel-held areas, then, was primarily coercive and unilateral in
nature and organised to ensure the military success of Renamo. Renamo’s strict
control over the formation and organisation of the auxiliaries reduced their role
to population control in a system of security governance that served rst and
foremost the rebel’s military agenda.
Auxiliary Armed Forces in Government-held Zones
Governance was dierent in Frelimo-held rural areas, where the political and
military structure became increasingly decentralised and disorganised in light
of Renamo’s threat. That resulted in the emergence of grassroots movements
for people’s self-defense, which, when aligning with Frelimo, enjoyed a higher
level of autonomy by comparison to the mujeeba.
The most important auxiliary armed forces on the government-side,34 the
Naparama, formed in the late 1980s in the border region between Zambézia
and Nampula after a large Frelimo counter-oensive failed to stop Renamo
advances and restore stability. The force reached, between 1988 and 1992, a
size of several thousand combatants.35 Renamo entered Zambézia in August
1982 (Legrand 1993, pp. 91, 92) and Nampula in April 1983 during a second
oensive across Zambézia (Do Rosário 2009, p. 305). Both provinces experienced
an escalation of violence in late 1986 when Renamo joined a local opposition
force and occupied several district towns along the border to Malawi and the
provincial border between Zambézia and Nampula. Frelimo feared that the
rebels would take control over the north and cut the country in half along the
Zambezi valley, the southern border of Zambézia province (Finnegan 1989, p.
62). In late 1986 and early 1987, the Mozambican military, together with allied
forces from Zimbabwe and Tanzania, began a counter-oensive. This operation
returned all district towns to Frelimo control by July 1988, but did not create
enduring stability.
Naparama’s emergence and diusion was fuelled by Frelimo’s inability to
protect the population from violence and displacement (Jentzsch 2014). The
main Naparama leader in Zambézia, the traditional healer Manuel António,
oered people an eective way to defend themselves.36 He claimed that he
had received a divine mission from Jesus Christ to liberate the Mozambican
people from the suering of the war and learned of a medicine to turn bullets
into water. He mobilised followers by ‘vaccinating’ them with this medicine.37
The eectiveness of the medicine depended on the respect of certain rules. For
example, Naparama combatants were
not allowed to look back, only look ahead; no one was allowed to be in front of the
other; no ghting in the shade, always in the sun; if the enemy was in the shade,
we were not allowed to be in the shade as well; (…) we could not retreat when
we heard shots, we had to go there where they [Renamo] were.38
All deaths amongst the Naparamas were explained by pointing to rule
violations.

CIVIL WARS 335
Learning from Renamo’s references to spiritual power, Naparama exploited
Renamo combatants’ belief systems and formed an innovative, collective
response to the rebel threat (Wilson 1992). By continuously advancing, often
while singing, Naparama created such fear amongst Renamo combatants that
it rarely came to a direct confrontation between the two forces. Renamo com-
batants ed as soon as they heard Naparamas approaching.
Although Frelimo was at rst sceptical, it soon co-opted the Naparama.
Worried by events in war-torn Angola, where two insurgent movements were
challenging the government, Frelimo ocials feared that Naparama would
evolve into a second insurgent force.39 Moreover, the local administration was
concerned about Naparama’s potential demands for support and compensation
during or after the war. Before Manuel António could work in Mocuba district, for
example, he had to ensure the local administration that his goal was not money
or political power, but only the protection of the population.40
When local ocials realised that Naparama was loyal and could contribute to
removing the military stalemate, most of them supported Naparama’s recruit-
ment eorts and some even agreed to joint military operations. The decision to
co-opt was thus based on pragmatic calculations to further local power inter-
ests, and not a change in the ocial party ideology of scientic socialism that
despised anything ‘traditional’. Frelimo ocials on the provincial and national
level never ocially acknowledged the cooperation with Naparama, although
the party abandoned all references to Marxism-Leninism at its party congress
in 1989 and changed its attitudes towards traditional authorities in the early
1990s.41
This ambivalent stance towards the auxiliary forces – the need to outsource
policing and military tasks in order to multiply forces yet the hesitation to sup-
port a ‘traditional’ force – and the ad hoc character in which Frelimo collabo-
rated with Naparama created distinct conditions in which the group was able to
operate relatively autonomously. The group quickly diused across districts and
expanded its tasks from nightly patrols and food distribution to independent
military operations. This is in stark contrast to Renamo’s strict control over the
mujeeba and their functions within the indirect system of rule, which separated
them from the military forces and operations.
The Naparama’s primar y task was to support Frelimo local governance, which
‘increasingly came to rely on war tactics – military control – and the use of
community villages as counter-insurgency devices’ similar to governance in
Renamo-held areas (Buur and Kyed 2007, p. 109). The Naparama took a prom-
inent role in the ‘recuperation’ of people from Renamo-held areas and settling
them in Frelimo-held villages. In fact, some people joined the Naparama in
order to bring back family members who had been captured by Renamo.42
However, the resettlement of people often occurred coercively, and Naparama
forces made sure that people, once in Frelimo-held areas, would not ee and
defect to Renamo-held areas:

336 C. JENTZSCH
We brought people [to Frelimo-held areas] and handed them over to the [local]
government. In some cases, these people ed and returned to the bush. We had
to go and search for them again. If the same person was recuperated more than
four times, we had to harm her to intimidate her.43
A Naparama combatant in Mecubúri in Nampula province conrmed that
those who did not want to return with the Naparama to Frelimo-held areas were
killed.44 Some people stayed in Frelimo-held areas only because they feared
being killed if they attempted to return to the areas from where they were
displaced.45
However, in contrast to Renamo’s mujeeba, Naparama did not only serve
Frelimo, but, as a grassroots movement, the people as well, and was therefore
much more involved in protecting the population. For example, Naparama was
tasked with nightly patrols to warn of imminent attacks.46 They also gathered
information in surrounding areas, partly by collaborating with local hunters
that were discontent with Renamo’s treatment of the population.47 Naparama
also accompanied peasants to their elds outside the village for the collection
of food.48
Little oversight by the provincial government and an increasingly dicult
military situation for Frelimo facilitated the expansion of Naparama’s activities
over time from more defensive to more oensive tasks. In the district of Alto
Molócuè, Manuel António rst worked on improving road security by treating
bus passengers with his medicine for their own protection, and then formed
small groups to attack Renamo strongholds in the northern part of the district
(Pereira 1999a, p. 86). Respondents frequently reported that Naparama came to
replace the army. A former Naparama combatant in Nampula reported that after
a successful operation in Mecubúri, the armed forces stationed in that district
‘decided that they would hand over their weapons to us and stop going into
battle until the time when the war ended’.49
Overall, by tolerating Naparama’s activities, Frelimo hoped to overcome a
military stalemate that had emerged around 1988 and substitute a weak and
ineective national army from which many soldiers had deserted. As a grassroots
movement that built signicantly on local belief systems, however, it did not
only serve the government, but was also involved in protecting the population
under Frelimo control, which made its activities less arbitrary than the mujeeba’s
and created a form of security governance that was dierent from the coercive
and oppressive one in rebel-held areas.50
Tactical Adaptation and Innovations in Security Governance
The success of the Naparama forces did not last long. Soon they were challenged
by their opponent’s learning and ‘tactical adaptation’ (McAdam 1983, p. 736),
the adoption of Naparama’s mobilisational technique. In addition, the main

CIVIL WARS 337
Naparama leader, Manuel António, was killed in December 1991, which severely
weakened the Naparama in its main area of operation, Zambézia province.
Probably towards the end of 1990 and the beginning of 1991, Renamo lead-
ers in Nampula and Zambézia provinces realised that they had to respond to
Naparama with a similar force in order to regain supremacy on the battleeld.
They decided that collaboration with traditional healers for ritual protection
would be the most promising response, essentially creating their own Naparama
force. Leaders identied traditional healers capable of initiating members of
Renamo’s existing auxiliaries – the mujeeba – to empower them to successfully
respond to Naparama’s threat. Although Renamo’s strict control over the aux-
iliaries initially limited mujeebas’ autonomy, developments on the battleeld
required them to expand mujeeba tasks to respond to Naparama’s threat.
One of the rst anti-Naparama forces was formed in Nampula, where tradi-
tional healers in the Renamo base of Namilasse in Murrupula district treated
youths with a vaccine to ght against Naparama and the Frelimo army.51 The
vaccine supposedly had the same eect as Naparama’s medicine. The group,
known under the name Mutapassa, used spears, but also had a few rearms.
Mutapassa combatants followed similar rules as the Naparama regarding pro-
hibited food, but their main rule was that they were not allowed to shoot unless
others did rst.52
In Zambézia, another anti-Naparama force emerged in the early months of
1991, led by the traditional healer Mulelepea (or Mulelepeya/Malelepea), which
may even have led to the defection of some Naparama combatants (Legrand
1993, p. 103). Renamo leaders chose between several traditional healers when
struggling to respond to Frelimo’s counter-oensive and nally settled on one
whose powers resembled most closely those of Naparama initiators:
Frelimo had intensied their attacks here in Nauela. This was when Renamo hired
Mulelepea of Namixaxen. However, there was another man called Namukhotxen
of the area of Nanthupa. Renamo asked the latter how they could solve the crit-
ical situation provoked by the enemy, and Namukhotxen answered that he was
capable [to help] because he could transform himself into a lion and decimate the
enemies. Renamo didn’t accept his proposal. By contrast, Mulelepea said that he
would use magic, vaccinating the warriors so that Frelimo’s bullets didn’t penetrate
their bodies. He was authorised to recruit men, usually youths called ‘anamavaka’
[spear users], to be vaccinated. They began their military operations. Renamo’s
guerrilla ghters advanced in the second line and the ‘anamavaka’ in the rst line
of oense.53
Mulelepea was an elder of about 70years and claimed that he had learned
how to transform himself into a child to escape detention by the cipaios
(Portuguese colonial police). When Mulelepea heard that there were Naparama
in Nauela, he claimed that Manuel António’s vaccine was weak and went to
Nauela to put his abilities into practice (Pereira 1999a, p. 95). He traveled to
other bases in other districts and also reached the regional base Maquiringa in
Namarrói district.54

338 C. JENTZSCH
In contrast to Naparama’s largely voluntary mobilisation, most of the
Mulelepea combatants were forcefully recruited from amongst the mujeeba.
In other regards, however, Mulelepea’s forces resembled those of Naparama. The
combatants had to follow similar rules of conduct as the Naparama.
55
Moreover,
similar to the way in which Naparama conducted joint operations with Frelimo,
the Mulelepea combatants advanced in front of the regular Renamo units to
make use of their special forces to clear the area.
Mulelepea’s warriors succeeded in posing a signicant threat to Naparama
and Frelimo units. The forces confronted each other for the rst time in April
1991 in Nauela, during which 25 Naparama combatants died (Pereira 1999a,
p. 98). Mulelepea’s followers were called ‘Khonkos’, which means the strong
and powerful (Pereira 1999a, p. 98). A former Frelimo soldier in Murrupula
claimed that ‘Khonkos’ denotes people who are ‘crazy’ and ‘don’t like to joke
around’.56 This demonstrates that Frelimo soldiers considered these forces as
unpredictable, which increased their threat on the battleeld. Naparama were
more afraid of the Khonkos than of Renamo combatants, as they claimed that
Mulelepea’s forces were dicult to catch with bare hands and – in contrast to
regular Renamo ghters – were able to kill Naparama with their spears (Pereira
1999a, p. 100).
In sum, Renamo’s response to Naparama brought new momentum to the
war in Zambézia province, after Frelimo seemed to have gained the military
advantage. Similar to Frelimo’s initial learning process that led to its (indirect)
support of the Naparama, Renamo was inuenced by failures on the battleeld
that gave the other side a military advantage. Tactical adaptation led to an
expansion of mujeebas’ tasks, who began to accompany Renamo combatants on
military operations. Although Renamo exploited local belief systems by forming
an ‘anti-Naparama’ force, however, the strict logic of strategic delegation – rather
than popular participation – in the formation of this new force resulted in the
same violent and arbitrary eect on security governance as the prior mujeeba
force.
Conclusion
The article has argued for a revision of the concept of armed group governance
to account for the multi-actor and multi-layered nature of security governance.
Security governance in civil war is characterised by the fragmentation of armed
actors and delegation of responsibilities. The formation of auxiliary armed forces
such as self-defense forces, militias and local patrols, is not only a ‘tool’ for state
actors to outsource violence and benet from local intelligence, but also for
rebel groups, shaping the senior partner’s relationship with civilians. When
resources are strained in a long and intense war, fragmentation of authority
occurs, as armed groups in power delegate tasks to other actors.

CIVIL WARS 339
However, security governance may dier between rebel-held and govern-
ment-held zones. These dierences have consequences for the relationship
between those in power and the local population. As the evidence has demon-
strated, the less involved civilians are in the formation of the auxiliaries, the less
autonomy an auxiliary has to shape its activities.
The rebel auxiliary forces in Mozambique were formed by Renamo and largely
focused on population control and resource extraction, which supported the
group’s focus on military objectives. In contrast, the auxiliary armed forces in
government-held areas, which were characterised by decentralised centres of
power, formed as a grassroots initiative and were more autonomous in expand-
ing their activities over time, in some cases even substituting for the state army.
This had distinct consequences for relationships with civilians. Civilians in rebel-
held areas perceived the auxiliaries and the system of governance they enforced
as exploitative and violent. In contrast, civilians in government-held areas were
actively involved in the formation of the auxiliaries, and so the protection of the
local population was a primary function of these new forces. This shows that
Renamo outsourced much of the violence to auxiliaries and considered them
to full a distinct function within their system of governance. For Frelimo, the
auxiliaries complemented their eorts, but did not replace repressive functions
of the state (Bertelsen 2016, Macamo 2016).
However, developments on the battleeld can lead to learning and innova-
tion in security governance, and an expansion of tasks. Both the Frelimo govern-
ment and the Renamo rebels learned from each other and adapted their tactics
when they realised that their prior tactics did not provide them with any military
advantage. Creating, supporting or adapting auxiliary armed forces was a way to
gain that advantage. As a consequence, security governance further militarised.
This article has made three main contributions. First, it demonstrates that
auxiliary armed forces contribute signicantly to armed group governance
and shape the interactions between those in power and the local population.
Second, the evidence highlights that governance in government-held areas
evolves during wartime in important ways. Too often, studies of political order
during wartime are focused only on rebel-held areas and do not suciently
study how governance in government-held areas is aected by the war. Lastly,
it has shown that in Mozambique, rebel and government forces used the same
governance ‘tools’ in the form of auxiliaries, but that their implementation was
characterised by important dierences.
Future research should explore further the role of auxiliary armed forces
in other cases in order to facilitate comparisons and conclusions about how
widespread the formation and uses of auxiliaries are and what eect they have
on armed group governance during civil war.

340 C. JENTZSCH
Notes
1. See the Introduction to this Special Issue.
2. Fieldwork for this project was conducted in 2010, 2011, 2012, and 2016 in the
capital Maputo, the provincial capitals Nampula and Quelimane, Mecubúri and
Murrupula districts in Nampula province and Lugela, Namarrói, and Nicoadala
districts in Zambézia.
3. The term comes from McAdam’s analysis of tactical innovation and adaptation
during the civil rights struggle in the USA in the 1950s and 1960s. The formation
of auxiliary armed forces could be understood as a tactical adaptation in civil war
that responds to tactical innovations on the opponent’s side. See McAdam (1983).
4. For an overview of the dierences of wartime developments across regions, see
the introduction in Englund (2002) and Cahen et al. (forthcoming). For a history
of the war, see Robinson (2006) and Emerson (2014).
5. For a critical re-evaluation of Renamo violence, see Schafer (2007). On the role
of violence in Frelimo state formation and during war, see Macamo (2016) and
Bertelsen (2016).
6. Interview with male resident (2011-11-28-m22, Murrupula-Chinga, Nampula).
In order to protect the identities of the respondents, the interview citations
indicate date, location, the interviewee’s role during the war, and gender of the
interviewees: N (Naparama); F (Frelimo combatant); R (Renamo combatant); M
(militiaman); P (religious leader); L (local leader including traditional and other
community leaders); H (traditional healer); G (government representative); m
(male); f (female). Frelimo had abandoned the system of traditional authorities
immediately after independence, accusing traditional leaders of having
collaborated with the colonial state. Frelimo sought to abandon what it called
‘obscurantism’ in society – all types of religion including traditional religion and
traditional healing. Alexander (1997) shows though that traditional authorities
continued to inuence local politics after independence.
7. There is an intense debate over the origins of the war in Mozambique and the
extent of popular support for Renamo. While scholars largely agree that without
external aggression from Rhodesia and South Africa, the war might not have
happened, domestic discontent and local conicts increased support for Renamo
and fuelled the war. See, for example, Roesch (1989), Minter (1994), Cahen (2000),
Lubkemann (2008, Chapter 3) and Jentzsch (2014).
8. 4334 Renamo soldiers (19.7 per cent of total ex-Renamo ghters) and 3073
government soldiers were aged between 10 and 14 at the time of their abduction
and can be considered child soldiers (Barnes 1997, p. 17).
9. Former Renamo combatants made frequent references to radios in bases and
advance posts in interviews with me.
10. The structure below the regional commanders is not as clear as the higher
command and there might have been regional variations (Vines 1991, p. 81).
11. Interview with former Renamo combatants (2012-03-08-Rm14, Nicoadala,
Zambézia; 2011-10-15-Rm2, Mecubúri, Nampula).
12. Interview with former Renamo combatant (2011-10-23-Rm3, Mecubúri, Nampula).
13. Interview with male resident (2011-11-28-m22, Murrupula, Nampula).
14. Geray (1990, p. 120 fn. 3) notes that the term mambo probably originated in
the Ndau region in central Mozambique, where most of Renamo’s ocers come
from. The origin of the term ‘mujeeba’ seems to come from the Zimbabwean
liberation struggle. Renamo learned the role and function of mujeebas from
the Zimbabwean National Liberation Army (ZANLA), as the Rhodesians trained

CIVIL WARS 341
early Renamo members in ZANLA methods and techniques (Wilson 1992, p. 541).
While the mujeeba were most prominent in the central region, across the border
from Zimbabwe and the heartland of Renamo activity, evidence from secondary
sources and my own interviews demonstrate the crucial role of mujeebas in
Zambézia and Nampula as well.
15. Interview with former mambo (2011-10-23-m7, Mecubúri, Nampula).
16. I have not seen this term used in other research on the war and my own sources
have also not referred to such a category of local ocials. This is probably due
to regional variations in systems of Renamo governance.
17. Interview with former Renamo civilian administrator (2012-06-22-Rm20, Regone,
Zambézia).
18. Interview with former Frelimo combatant (2011-09-13-Fm1, Nicoadala, Zambézia).
19. Interview with former Naparama combatant (2011-09-09-Nm4, Nicoadala,
Zambézia).
20. Interview with former Renamo combatant (2012-03-08-Rm14, Nicoadala,
Zambézia). Vines (1991, pp. 92, 93) notes that this version of a base is a regional
variation of Renamo’s bases in the north. In southern regions, where bases were
more mobile, combatants lived farther away from the population.
21. Interviews with former Renamo combatant (2011-10-23-Rm3, Mecubúri,
Nampula) and former Renamo civilian administrator (2012-06-22-Rm20, Regone,
Zambézia).
22. Interview with former Renamo combatant (2011-08-18-Rm1, Maputo, Zambézia).
23. Interview with former Renamo combatant (2011-10-23-Rm3, Mecubúri, Nampula).
24. Interviews with a woman and a man who lived in Renamo-controlled areas
(2011-11-25-f11, Murrupula-Nampaua; 2011-11-24-m18, Murrupula-Namilasse,
Nampula).
25. Interview with Renamo Secretary General Fernandes in 1985, quoted in Hall
(1990, p. 50).
26. Interview with former Renamo civilian administrator (2012-06-22-Rm20, Regone,
Zambézia).
27. Also see Pereira (1999a, p. 46). This was again similar to ZANLA’s mujeeba, who
often punished people for not respecting and listening to the guerrillas (Maxwell
1993, p. 374).
28. Some scholars assert that mujeebas were recruited amongst former cipaios
(‘native police during the colonial era’), which would have provided them with
experience in policing and intelligence (Hall 1990, p. 56).
29. Interview with a group of community leaders who lived under Renamo control
(2012-06-14-Gr-Lm4, Lugela-Tacuane, Zambézia).
30. Interview with community leader (2011-10-17-Lm11, Mecuburi, Nampula).
31. Interview with community leader (2011-10-17-Lm11, Mecuburi, Nampula);
Interview with local government ocial (2011-11-10-Gm13, Nampula); Interview
with former Naparama combatant (2011-11-04-Nm37, Murrupula, Nampula);
Interview with male resident (2011-11-06-m11, Murrupula, Nampula). Also see
Hall and Young (1997).
32. Interview with community leader (2011-09-23-Lm3a, Nicoadala, Zambézia).
33. Frelimo’s suspicion reected a more general sentiment that the mujeeba were
easily corrupted and often worked for both sides of the conict. Interview with
community leader (2011-09-21-Lm1, Nicoadala, Zambézia).
34. After independence, Frelimo mobilised civilians into ‘popular militias’. Once
Renamo became a severe threat in the early 1980s, militias were tasked with
vigilance. These militias, lacking in supplies and morale, did not manage to build

342 C. JENTZSCH
a successful defense against Renamo attacks in Zambézia and Nampula and so
community residents developed their own forms of protection. See Jentzsch
(2014).
35. The number of Naparama combatants is dicult to ascertain. Its main leader in
Zambézia, Manuel António, claimed to have about 14,000 ghters in May 1991
(Waterhouse 1991, p. 14). However, the journalist Gil Lauriciano, who covered the
war in Zambézia extensively, estimates that the group did not have more than
2000 members (personal communication, July 2010). Based on my interviews
with former Naparama members that indicate that many districts had about
200 Naparama, which only included those in the main district town, I estimate
the size to about 4000–6000 members across both provinces, Zambézia and
Nampula. The current Naparama leadership claims to have registered 4438 former
Naparama in four districts, which are Inhassunge, Nicoadala, Namacurra and
Mopeia (Interview 2011-08-23-Gr-Nm1, Quelimane, Zambézia). As a comparison,
Renamo was estimated to have about 20,000 combatants.
36. Interviews with former Naparama combatants (2011-09-19-Nm11, Nicoadala,
Zambézia; 2011-09-20-Fm2-N, Nicoadala, Zambézia). For a detailed history of
the Naparama and its leadership, see Jentzsch (2014).
37. Interview with former Naparama combatant (2011-09-09-Nm2, Nicoadala,
Zambézia).
38. Interview with former Naparama combatant (2011-09-09-Nm2, Nicoadala,
Zambézia).
39. Interview with provincial government representative (2011-10-10-Gm7,
Nampula).
40. Interview with former Naparama leader (2012-06-06-Nm46, Lugela, Zambézia).
41. Due to this strict denial of Frelimo-Naparama cooperation on the national
level, Naparama was not considered a party to the conict during the peace
negotiations between Frelimo and Renamo and was therefore not included in
post-war demobilisation programs (Coelho and Vines 1992).
42. Six respondents in Nicoadala, Zambézia, said they joined to bring back family
members and friends. E.g., Interview with former Naparama combatant (2011-
09-09-Nm4, Nicoadala, Zambézia).
43. Interview with former Naparama combatant (2011-09-14-Nm10, Nicoadala,
Zambézia).
44. Interview with Naparama combatant (2011-10-16-Nm25, Mecubúri, Nampula).
45. Interview with female resident (2011-11-25-f10, Murrupula-Nampaua, Nampula).
46. Interview with local leader (2011-09-23-Lm3a, Nicoadala, Zambézia); Interview
with female resident (2011-09-26-Lf1, Nicoadala, Zambézia); Interview with
Naparama commander (2012-06-10-Nm46, Lugela, Zambézia); Interview with
male resident (2011-10-02-m5, Nicoadala, Zambézia).
47. Interview with Naparama combatant (2011-09-14-Nm10, Nicoadala, Zambézia).
48. Interview with Naparama commander (2012-06-10-Nm46, Lugela, Zambézia).
49. Interview with former Naparama combatant (2011-10-22-Nm27, Mecubúri,
Nampula).
50. There is some regional and temporal variation in Naparama’s perpetration of
violence, which I cannot discuss here due to space constraints, but as Naparama
became more successful over time, it seemed to also become more violent
against civilians, especially in Renamo-held areas. See Jentzsch (2014).
51. Interview with male resident (2011-11-24-m18, Namilasse, Murrupula, Nampula).
Other interviewees identied the traditional healer as Sabala from Taveia in
Ribáuè district in Nampula who had treated youths in Ribáuè-Sede. See Interviews

CIVIL WARS 343
with male resident (2011-11-28-m22, Chinga, Murrupula, Nampula) and former
Frelimo combatant (2011-11-28-Fm13, Nampaua, Murrupula, Nampula).
52. Interviews with male residents (2011-11-24-m18; 2011-11-24-m19, Namilasse,
Murrupula, Nampula).
53. Interviewee cited in Pereira (1999a, p. 94). Translation from Portuguese by the
author.
54. Interview with former Renamo combatants (2012-06-22-Gr-Rm3, Rumala,
Namarrói, Zambézia).
55. Interview with former Renamo combatants (2012-06-22-Gr-Rm3, Rumala,
Namarrói, Zambézia).
56. Interview with former Frelimo combatant (2011-11-28-Fm13, Nampaua,
Murrupula, Nampula).
Acknowledgments
I thank the editors of this special issue and participants at workshops at Oxford University
and the University of Amsterdam, the editors of this journal and two anonymous referees
for helpful suggestions on earlier drafts of this article.
Disclosure Statement
No potential conict of interest was reported by the author.
Funding
This work was supported by a US National Science Foundation Doctoral Dissertation
Research Improvement Grant [grant number 1123960], a Yale MacMillan Center
International Dissertation Research Grant, and a Netherlands Organisation for Scientic
Research (NWO) Veni Grant [grant number 451-16-008].
Notes on Contributor
Corinna Jentzsch is assistant professor of international relations at Leiden
University in the Netherlands. Her research focuses on civil wars and the emer-
gence of informal institutions of security governance, community mobilisa-
tion against insurgent violence in Mozambique and conditions of successful
peacekeeping in African conicts. Her work has been published in the Journal
of Conict Resolution and the African Conict and Peacebuilding Review, and she
is a contributor to Africa is a Country.
ORCID
Corinna Jentzsch http://orcid.org/0000-0003-0936-6984

344 C. JENTZSCH
References
Abrahamsen, R. and Williams, M.C., 2008. Public/private, global/local: the changing
contours of Africa’s security governance. Review of African Political Economy, 35 (118),
539–553.
Alexander, J., 1997. The local state in post-war Mozambique. Political practice and ideas
about authority. Africa: Journal of the International African Institute, 67 (1), 1–26.
Arjona, A., 2014. Wartime institutions: a research agenda. Journal of Conict Resolution,
58 (8), 1360–1389.
Arjona, A., 2016. Rebelocracy. Social order in the Colombian civil war. Cambridge, MA:
Cambridge University Press.
Arjona, A., Kasr, N., and Mampilly, Z.C., eds., 2015. Rebel governance in civil war.
Cambridge, MA: Cambridge University Press.
Bakke, K.M., Cunningham, K.G., and Seymour, L.J.M., 2012. A plague of initials:
Fragmentation, cohesion, and inghting in civil wars. Perspectives on Politics, 10 (2),
265–283.
Barnes, S., 1997. The socio-economic reintegration of demobilised soldiers in Mozambique.
The soldiers’ view. UNDP. Maputo: United Nations Development Program.
Bertelsen, B.E., 2016. Violent becomings. State formation, sociality, and power in
Mozambique. New York: Berghahn.
Biberman, Y., 2016. Self-defense militias, death squads, and state outsourcing of violence
in India and Turkey. Journal of Strategic Studies, online rst.
Breidlid, I.M. and Arensen, M.J., 2014. ‘Anyone who can carry a gun can go:’ the role of the
White Army in the current conict in South Sudan. PRIO Paper. Oslo: PRIO.
Buur, L. and Kyed, H.M., 2007. Traditional authority in Mozambique: the legible space
between state and community. In: L. Buur and H.M. Kyed, eds. State recognition and
democratization in Sub-Saharan Africa. A new dawn for traditional authorities. New York:
Palgrave Macmillan, 105–127.
Cahen, M., 2000. Nationalism and ethnicities: lessons from Mozambique. In: E. Braathen,
M. Bøås, and G. Sæther, eds. Ethnicity kills? The politics of war, peace and ethnicity in
Sub-Saharan Africa. New York: St. Martin’s Press, 163–187.
Cahen, M., Morier-Genoud, E., and Do Rosário, D., eds., Forthcoming. The war within: new
perspectives on the civil war in Mozambique, 1976–1992. Suolk: James Currey.
Carey, S., Mitchell, N.J., and Lowe, W., 2013. States, the security sector and the monopoly
of violence: a new database on pro-government militias. Journal of Peace Research,
50 (2), 249–258.
Christia, F., 2012. Alliance formation in civil wars. New York: Cambridge University Press.
Clayton, G. and Thomson, A., 2014. The enemy of my enemy is my friend… The dynamics
of self-defense forces in irregular war: the case of the sons of Iraq. Studies in Conict
& Terrorism, 37 (11), 920–935.
Coelho, J.P.B., 1998. State resettlement policies in post‐colonial rural Mozambique: the
impact of the communal village programme on Tete province, 1977–1982. Journal of
Southern African Studies, 24 (1), 61–91.
Coelho, J.P.B. and Vines, A., 1992. Pilot study on demobilization and re-integration of ex-
combatants in Mozambique, USAID’s Mozambique Demobilization and Reintegration
Support Project (656–0235). University of Oxford, Refugee Studies Programme, Queen
Elizabeth House. Oxford: University of Oxford.
De Waal, A., 2004. Counter-insurgency on the cheap. Review of African Political Economy,
31 (102), 716–725.

CIVIL WARS 345
Dieterich, A., 2016. Are FARC militias a wild card in Colombia peace process? InSight Crime
[online]. Available from: http:\\www.insightcrime.org/news-briefs/are-farc-militias-a-
wild-card-in-colombia-peace-process [Accessed 1 February 2017].
Do Rosário, D., 2009. Les Mairies des « autres » : une analyse politique, socio-historique et
culturelle des trajectoires locales. Le cas d’Angoche, de l’Île de Moçambique et de Nacala
Porto [The Municipalities of ‘Others’. The Political, Socio-historical and Cultural Analysis
of the Local Trajectories: A Case Study of Angoche, Mozambique Island and Nacala
Porto]. Thesis (PhD). Université Montesquieu/Bordeaux IV.
Dobbin, F., Simmons, B., and Garrett, G., 2007. The global diusion of public policies:
social construction, coercion, competition, or learning? Annual Review of Sociology,
33 (1), 449–472.
Emerson, S.A., 2014. The battle for Mozambique. The FRELIMO-RENAMO struggle, 1977–
1992. Pinetown: 30 Degrees South Publishers.
Englund, H., 2002. From war to peace on the Mozambique-Malawi borderland. Edinburgh:
Edinburgh University Press for the International African Institute, London.
Finnegan, W., 1989. A reporter at large. I-The emergency. The New Yorker (May 22), 43–76.
Finnegan, W., 1992. A complicated war: the harrowing of Mozambique. Berkeley: University
of California Press.
Geray, C., 1990. La cause des armes au Mozambique. Anthropologie d’une guerre civile
[The Origin of War in Mozambique. An Anthropology of a Civil War]. Paris: Karthala.
Gersony, R., 1988. Summary of Mozambican refugee accounts of principally conict-related
experience in Mozambique. Bureau for Refugee Programs. Washington DC: Department
of State.
Hall, M., 1990. The Mozambican National Resistance Movement (Renamo): a study in
the destruction of an African country. Africa: The Journal of the International African
Institute, 60 (1), 39–68.
Hall, M. and Young, T., 1997. Confronting leviathan: Mozambique since independence.
Athens: Ohio University Press.
Hanlon, J., 1984. Mozambique. The revolution under re. London: Zed Books.
Homan, D., 2011. The war machines: young men and violence in Sierra Leone and Liberia.
Durham, NC: Duke University Press.
Huang, R., 2016. The wartime origins of postwar democratization: civil war, rebel governance,
and political regimes. Cambridge, MA: Cambridge University Press.
Jackson, P., 2003. Warlords as alternative forms of governance. Small Wars and Insurgencies,
14 (2), 131–150.
Jentzsch, C., 2014. Militias and the dynamics of civil war. Thesis (PhD). Yale University.
Jentzsch, C., Kalyvas, S.N., and Schubiger, L.I., 2015. Militias in civil wars. Journal of Conict
Resolution, 59 (5), 755–769.
Kalyvas, S.N., 2006. The logic of violence in civil war. Cambridge, MA: Cambridge University
Press.
Kasr, N., 2015. Rebel governance. Constructing a eld of inquiry: denitions, scope,
patterns, order, causes. In: A. Arjona, N. Kasr, and Z.C. Mampilly, eds. Rebel governance
in civil war. Cambridge, MA: Cambridge University Press, 21–46.
Krahmann, E., 2003. Conceptualizing security governance. Cooperation and Conict, 38
(1), 5–26.
Legrand, J.-C., 1993. Logique de guerre et dynamique de la violence en Zambézie,
1976–1991 [Logic of War and Dynamics of Violence in Zambézia, 1976–1991]. Politique
Africaine, June (50), 88–104.
Lemia, R.J., 2001. Pós-Independência, guerra e reassentamento da população no distrito
de Namacurra (1975–1998/9) [Post-independence, War and Return of the Population

346 C. JENTZSCH
in Namacurra District (1975–1998/9]. Thesis (Licenciatura). Universidade de Eduardo
Mondlane.
Lubkemann, S.C., 2005. Migratory coping in wartime Mozambique: an anthropology
of violence and displacement in ‘fragmented wars’. Journal of Peace Research, 42 (4),
493–508.
Lubkemann, S.C., 2008. Culture in chaos. An Anthropology of the social condition in war.
Chicago, IL: University of Chicago Press.
Macamo, E., 2016. Violence and political culture in Mozambique. Social Dynamics, 42
(1), 85–105.
Malejacq, R., 2016. Warlords, intervention, & state consolidation: a typology of political
orders in failed & failing states. Security Studies, 25 (1), 85–110.
Mampilly, Z.C., 2011. Rebel rulers. Insurgent governance and civilian life during war. Ithaca,
NY: Cornell University Press.
Maxwell, D.J., 1993. Local politics and the war of liberation in North-East Zimbabwe.
Journal of Southern African Studies, 19 (3), 361–386.
McAdam, D., 1983. Tactical innovation and the pace of insurgency. American Sociological
Review, 48 (6), 735–754.
Menkhaus, K., 2007. Governance without government in Somalia: Spoilers, state building,
and the politics of coping. International Security, 31 (3), 74–106.
Minter, W., 1994. Apartheid’s contras: An inquiry into the roots of war in Angola and
Mozambique. Johannesburg, London / Atlantic Highlands, NJ: Witwatersrand University
Press / Zed Books.
Mitchell, N.J., Carey, S., and Butler, C., 2014. The impact of pro-government militias on
human rights violations. International Interactions, 40 (5), 812–836.
Peic, G., 2014. Civilian defense forces, state capacity, and government victory in
counterinsurgency wars. Studies in Conict & Terrorism, 37 (2), 162–184.
Pereira, F.M., 1999a. Particularidades da dinâmica do conito armado no distrito do Alto
Molócuè, 1982–1992: Violência armada e guerra mágica [Peculiarity of the Dynamics of
Armed Conict in Alto Molócuè District, 1982–1992: Armed Violence and Magic War].
Thesis (Licenciatura). Universidade Eduardo Mondlane.
Pereira, J.C.G., 1999b. The politics of survival. Peasants, chiefs and Renamo in Maringue
district, Mozambique 1982–1992. Thesis (MA). University of the Witwatersrand.
Reno, W., 1998. Warlord politics and African states. Boulder, CO: Lynne Rienner.
Richards, J., 2014. An institutional history of the Liberation Tigers of Tamil Eelam (LTTE).
CCDP Working Paper, 10. Geneva: The Graduate Institute, The Centre on Conict,
Development and Peacebuilding.
Robinson, D.A., 2006. Curse on the land: a history of the Mozambican civil war. Thesis (PhD).
University of Western Australia.
Roesch, O., 1989. Is Renamo a popular Movement in Mozambique? Southern African
Review of Books, 12 (December/January), 20–22.
Roesch, O., 1992. RENAMO and the peasantry in southern Mozambique: a view from Gaza
Province. Canadian Journal of African Studies, 26 (3), 462–484.
Schafer, J., 2007. Soldiers at peace. Veterans and society after the civil war in Mozambique.
New York: Palgrave Macmillan.
Schubiger, L.I., 2013. Repression and mobilization in civil war. The consequences of state
violence for wartime collective action. Thesis (PhD). University of Zurich.
Seymour, L.J.M., 2014. Why factions switch sides in civil wars. Rivalry, patronage, and
realignment in Sudan. International Security, 39 (2), 92–131.
Sinno, Abdulkader H., 2008. Organizations at war in Afghanistan and beyond. Ithaca, NY:
Cornell University Press.

CIVIL WARS 347
Staniland, P., 2012. States, insurgents, and wartime political orders. Perspectives on Politics,
10 (2), 243–264.
Staniland, P., 2017. Armed politics and the study of intrastate conict. Journal of Peace
Research, 54 (4), 459–467.
Velasquez, H.A., 1995. Milicias: La guerrilla en las ciudades [Militias: The Guerilla in the
Cities]. El Tiempo [online], 15 October. Available from: http://www.eltiempo.com/
archivo/documento/MAM-423979 [Accessed 1 February 2017].
Vines, A., 1991. Renamo: terrorism in Mozambique. London: Centre for Southern African
Studies, University of York in association with James Currey.
Waterhouse, R., 1991. Antonio’s triumph of the spirits. Africa South (Harare), May, 14–15.
Weinstein, J.M., 2007. Inside rebellion. The politics of insurgent violence. Cambridge, MA:
Cambridge University Press.
Wiegink, N., 2013. Why did the soldiers not go home? Demobilized combatants, family
life, and witchcraft in postwar Mozambique. Anthropological Quarterly, 86 (1), 107–132.
Wilson, K.B., 1992. Cults of violence and counter-violence in Mozambique. Journal of
Southern African Studies, 18 (3), 527–582.
Wood, E.J., 2008. The social processes of civil war: the wartime transformation of social
networks. Annual Review Political Science, 11, 539–561.